Tournament compensation systems are widely used in practice and have been extensively analyzed theoretically. However, one major problem has hardly been studied in a company context so far: Although it is theoretically well understood that tournament compensation systems are only effective when employees are homogeneous, it has rarely been analyzed what companies can do when they are confronted with employee heterogeneity. In our paper we derive hypotheses on the performance effects of tournament compensation systems in a context of employee heterogeneity based on tournament and expectancy theory. Using personnel records from incentive travel contests we are able to show that performance is lower in a situation with employee heterogeneity, but that in heterogeneous tournaments, incentives may still work for a sub-group of employees whom we term the "threshold group". In addition, we also show how companies manage to design the information disclosure systems necessary to install effective "handicapping" or "league-building" systems in order to increase the effort-performance-expectancy.